See how an artificial brain makes decisions. Start with one neuron and turn its knobs, then build a small network and watch it learn. How it works →
Horizontal: the sum z. Vertical: the neuron's output.
The shaded background is the network's current guess for every spot. Dots with a dark ring are still wrong. Click the board to add your own points.
Loss is the network's score for how wrong it is. Training pushes it down.
Each square shows what one neuron "sees" across the board: blue where it fires strongly, orange where it stays low. Lines show weights, thicker means stronger, blue is positive, orange is negative. With no hidden layers the network is a single neuron.